Compute-forward multiple access (CFMA) with nested LDPC codes
Author(s)
Erixhen Sula; Jingge Zhu; Adriano Pastore; 임성훈; Michael Gastpar
Alternative Author(s)
임성훈
Publication Year
2017-06-30
Abstract
Inspired by the compute-and-forward scheme fromNazer and Gastpar, a novel multiple-access scheme introducedby Zhu and Gastpar makes use of nested lattice codes andsequential decoding of linear combinations of codewords torecover the individual messages. This strategy, coined compute-forward multiple access (CFMA), provably achieves points onthe dominant face of the multiple-access capacity region whilecircumventing the need of time sharing or rate splitting. For atwo-user multiple-access channel (MAC), we propose a practicalprocedure to design suitable codes from off-the-shelf LDPCcodes and present a sequential belief propagation decoder withcomplexity comparable with that of point-to-point decoders. Wedemonstrate the potential of our strategy by comparing severalnumerical evaluations with theoretical limits. individual messages. This strategy, coined compute-forward multiple access (CFMA), provably achieves points onthe dominant face of the multiple-access capacity region whilecircumventing the need of time sharing or rate splitting. For atwo-user multiple-access channel (MAC), we propose a practicalprocedure to design suitable codes from off-the-shelf LDPCcodes and present a sequential belief propagation decoder withcomplexity comparable with that of point-to-point decoders. Wedemonstrate the potential of our strategy by comparing severalnumerical evaluations with theoretical limits.